A secure email gateway is a filter that sits between the internet and your mailboxes. Every message bound for your firm passes through it first. It checks who sent the message, scans links and attachments, and scores it for spam and phishing. Then it delivers the message, quarantines it, or turns it away. Most gateways check outgoing mail on the way out too.
If your firm uses Microsoft 365 or Google Workspace, you already have one of these. It came with the subscription. The real question for a small firm is whether to add a second layer, and which kind. This guide covers how a gateway works, what it catches and what it can’t, and how to decide.
How a secure email gateway handles each message
A gateway works because of one DNS setting. Your domain’s MX record tells the rest of the internet where to deliver your mail. Point it at a gateway, and every inbound message lands there first. The gateway then passes clean mail on to your mailbox provider.
Inside, the checks run in a set order. Microsoft documents its own order for the protection built into every Microsoft 365 cloud mailbox (Microsoft Learn, updated August 2026). Third-party gateways follow much the same shape:
- Connection filtering. The sending server’s reputation is checked. Microsoft says most spam is rejected right here.
- Malware scanning. Attachments and active content are scanned. A hit goes to quarantine.
- Your own rules. Any mail flow rules your firm has set run next, such as “flag anything from this domain.”
- Spam and phishing scoring. The message is graded as spam, bulk, phishing or high confidence phishing, and checked for spoofing.
- Delivery or hold. Clean mail goes to the inbox. The rest goes to junk or quarantine, depending on policy.
Paid tiers and third-party products add more on top. The common extras are rewriting links so they’re checked again when clicked, opening attachments in a sandbox before delivery, and flagging mail that imitates your own people or domain. On outbound mail, a gateway can encrypt messages or stop one that carries data it shouldn’t.
A held message waits in quarantine until a policy or a person decides. What that pile is and how it works is covered in what quarantine means in an inbox.
What it catches, and what gets past
A gateway is very good at volume. Known bad senders, known malware, bulk spam and lookalike pages that thousands of people received this week: that’s its home ground. Most of the junk you never see was stopped by a filter like this.
It’s weaker where the message looks ordinary. Four cases slip through most often:
- Mail from a real account that’s been taken over. The sender is genuine, the domain is genuine, and the thread is genuine. Nothing in the message is technically wrong. What to do when it’s a client’s account is in when a client’s email is taken over.
- A plain-text request with no link or file. “Please use our new bank details for this invoice” has nothing for a scanner to open. The check that stops it is a phone call, set out in the fake supplier invoice.
- A lookalike domain with its paperwork in order. Microsoft’s own documentation says impersonation can pass SPF, DKIM and DMARC if the attacker sets up valid records for a lookalike domain.
- Mail between your own people. A message from one colleague’s mailbox to another never crosses the MX record. A gateway in front of your domain doesn’t see it, unless the product adds a separate internal feed.
These are the costly ones. The FBI’s 2025 Internet Crime Report records 24,768 business email compromise complaints in 2025, with reported losses of $3,046,598,558. That’s up from 21,442 complaints and $2,770,151,146 in 2024. Most of that fraud rides on exactly the messages above.
So the filter takes most of the weight, and a few judgments stay with the person reading. How to make that judgment on a single message is in how to spot a phishing email.
Secure email gateway vs ICES
The newer alternative is called integrated cloud email security, or ICES. According to a Darktrace post from May 2023, Gartner coined the term in 2021. It describes tools that connect to Microsoft 365 or Google Workspace by API, with no change to your MX record.
Microsoft’s security blog of July 17, 2025 puts the difference in terms of order. A gateway filters mail before Microsoft’s protection sees it. An ICES tool runs after it, as a second filter, and moves messages it doesn’t like out of the inbox.
| Secure email gateway | ICES (API-based) | |
|---|---|---|
| Where it sits | In front of your mailbox provider | Inside your Microsoft 365 or Google tenant |
| How it connects | Your MX record points to it | An API connection your admin approves |
| When it acts | Before delivery | Mostly after delivery, by pulling mail back |
| Sees internal mail | Not by default | Usually, since it reads the mailboxes |
| Sees mail from a taken-over account | Only what crosses the gateway | Can compare it with that sender’s usual behavior |
| Outbound controls | Strong: encryption, data-loss rules, routing | Lighter, varies by product |
| Setup effort | DNS change and mail-flow testing | Usually lighter |
| If the vendor has an outage | Inbound mail can queue | Mail still arrives, unchecked by that layer |
The line between the two is blurring. Proofpoint’s 2025 decision guide sells the same product both ways. It says a gateway deployment can take up to a few weeks. Cloudflare’s documentation lists four ways to deploy its email security: API, BCC, journaling, or MX inline. So ask any vendor which mode they’re proposing for your firm.
Does Microsoft 365 offer a secure email gateway?
Yes. Every Microsoft 365 subscription with cloud mailboxes includes built-in protection against spam, malware, phishing and spoofing. Microsoft says it’s on by default and can’t be switched off, only tightened. When your MX record points to Microsoft 365, that protection is your gateway. Even a competitor calls it that: Proofpoint’s guide refers to it as “the Microsoft SEG.”
What changes from plan to plan is how much sits on top. Microsoft calls it a protection ladder (Microsoft Learn, updated May 2026):
| Layer | Who has it | What Microsoft says it adds |
|---|---|---|
| Built-in protection (once called Exchange Online Protection) | Every subscription with cloud mailboxes | Blocks “broad, volume-based, known email attacks”: spam, malware, spoofing, quarantine |
| Defender for Office 365 Plan 1 | Included in Microsoft 365 Business Premium and E3, or as an add-on | Impersonation protection for your people and domains, Safe Links, Safe Attachments, protection in Teams |
| Defender for Office 365 Plan 2 | Included in Microsoft 365 E5, or as an add-on | Phishing simulations, threat hunting, automated investigation and response |
For most small firms, the step that matters is Plan 1. Impersonation protection is the part that looks for “your managing partner” writing from a stranger’s domain. If you’re on Business Basic or Business Standard, you don’t have it unless you’ve added it. Your IT provider can tell you which plan you’re on in a minute.
Microsoft also recommends turning on its Standard or Strict preset security policies for everyone. They’re Microsoft’s tested settings, and they replace the defaults most tenants never touch. That’s a good question for whoever runs your Microsoft 365.
If you’re asking because of an email that claims to be from Microsoft, that’s a different question. The real Microsoft sender addresses, the Teams “someone is trying to reach you” notice and the “unusual sign-in activity” alert are all covered in checking phishing that poses as Microsoft.
Google Workspace has one too
Gmail’s filtering works the same way for Workspace accounts. Google’s Safety Center states it plainly:
“Gmail blocks more than 99.9% of spam, phishing attempts, and malware from reaching you.”
That’s Google’s figure about its own product, and no independent test sits behind it on that page. It does tell you the baseline is high. As with Microsoft, the gaps that remain are the four ordinary-looking messages above.
When a second gateway makes things harder
Putting a third-party gateway in front of Microsoft 365 or Gmail isn’t free, even before the invoice. The two filters now see the same mail from different angles.
Microsoft explains the core problem in its page on Enhanced Filtering for Connectors (updated August 2026). Once mail passes through a gateway, it reaches Microsoft carrying the gateway’s address in place of the original sender’s. Microsoft’s own sender checks then have less to go on. Microsoft offers a setting to recover the original details, but someone has to set it up and keep it right.
Google describes the same situation for Workspace in its help page on inbound mail gateways:
“Gmail does not enforce the sender’s DMARC policy for messages that arrive from the inbound gateway.”
In practice that means two quarantines to check, two sets of rules that can disagree, and an MX record your whole firm depends on. None of that is a reason to avoid a gateway. It’s a reason to have someone who owns it.
It’s also fair to know where the numbers come from. Microsoft’s July 2025 blog reported its own analysis that adding an ICES tool after Defender improved malicious-message catch by an average of 0.30%. That’s Microsoft measuring the market it competes in. Vendors selling extra layers publish figures pointing the other way. Neither is an independent test.
Who needs a separate secure email gateway
You probably don’t need one if most of these are true:
- All your mail lives in Microsoft 365 or Google Workspace.
- You have, or can add, Defender for Office 365 Plan 1 or the Google equivalent.
- Nobody on staff or on contract owns email security day to day.
- Your biggest worries are fake invoices and taken-over client accounts.
For that firm, the money usually does more on the subscription’s own top tier, on phishing-resistant sign-in, and on a payment rule people actually follow. A rule for confirming bank-detail changes is in the fake supplier invoice.
A separate gateway or ICES layer starts to earn its place when one of these is true:
- You still run Exchange on your own servers, or mail is split across more than one provider. A gateway gives you one front door for all of it.
- A client contract, insurer or regulator names a control your current plan doesn’t provide, such as outbound encryption or archiving with a specific retention.
- You need outbound rules that stop certain data leaving, beyond what your plan includes.
- An IT provider already manages one across its clients and will run it for you.
- You’ve been hit by mail your current filter passed, and a review shows a second opinion would have caught it.
If the last one is your reason, look at an ICES tool before a gateway. Those incidents usually involve the ordinary-looking messages that an API tool, watching sender behavior from inside the mailbox, is built for.
How to compare the leading options
There isn’t one best secure email gateway. There’s the one that fits how your mail flows and who’ll run it. Gartner ranks the field each year. Its Magic Quadrant for Email Security Platforms of December 16, 2024, is behind a paywall, but vendors often republish their placement.
Start with these questions, in this order:
| Question | Why it matters |
|---|---|
| Gateway, API, or both? | Decides whether you change your MX record, and what happens in an outage |
| Does it see internal mail? | Mail between your own people is how a taken-over account spreads |
| How does it judge a message with no link or file? | That’s the payment-redirect email, and the costly one |
| Who reviews quarantine, and how fast? | A held message from a client is a missed deadline if nobody looks |
| What does it add to what you already pay for? | Ask for its detections on your own mail during a trial, not a brochure number |
| Who configures and watches it? | A filter nobody owns drifts out of date |
| What does it cost per mailbox, with your current plan included? | Compare against stepping up your Microsoft or Google tier |
Some of the names you’ll meet, grouped by how they connect. Each description comes from the vendor’s own page or Microsoft’s:
| Option | How it connects | Notes |
|---|---|---|
| Microsoft Defender for Office 365 | Built into Microsoft 365 | Plan 1 in Business Premium; also sold as an add-on |
| Gmail’s built-in protection | Built into Google Workspace | On for every account |
| Proofpoint Core Email Protection | Gateway or API | Its own guide covers when to choose each |
| Mimecast | Gateway | Describes itself as a cloud-native gateway for Microsoft 365 or Google Workspace; internal mail needs a separate journal feed |
| Cloudflare Email Security | API, BCC, journaling or MX inline | Four deployment modes in its documentation |
| Darktrace/EMAIL, KnowBe4 Defend, VIPRE Integrated Email Security | API, after Microsoft’s filter | The three vendors in Microsoft’s ICES Vendor Ecosystem, which needs Defender Plan 2 or E5 |
Microsoft’s ecosystem matters for one practical reason. With a member vendor, its quarantine shows up in the same Defender screen as Microsoft’s own. That’s one place to look instead of two. For a small firm, fewer places to check is worth a lot.
After the filter: the inbox you read
Every option above works on the message before you see it. Point works at the next step, on the mail your provider delivers.
When a risky or malicious message gets through, Point moves it into a separate pile. You look at it when it suits you. A stranger’s mail is held as well, and Point takes no instructions from it. That’s a reading layer, and it sits behind whatever gateway you choose, not in place of it. How Point keeps hostile mail out of its own reading is in keeping suspicious mail away from the AI. What else it does for a small firm is on the benefits page.
The ordinary-looking message from a real but taken-over account still reaches a person. So the habits in how to spot a phishing email stay yours, whatever sits in front.
Common questions
What is a secure mail gateway?
It’s the same thing as a secure email gateway. It’s a service your mail passes through before it reaches your mailboxes. It checks the sender, scans links and attachments, and quarantines or blocks what looks dangerous. Most also check outgoing mail.
What are the best secure email gateways?
There’s no single best one. For a firm already on Microsoft 365, the first option to weigh is Defender for Office 365 Plan 1, which comes with Business Premium. Proofpoint and Mimecast are the long-standing gateway names. Darktrace, KnowBe4 and VIPRE plug into Microsoft’s own screen as API tools. Pick by how your mail flows, whether the tool sees internal mail, and who’ll run it.
Does Microsoft 365 offer a secure email gateway?
Yes. Every Microsoft 365 plan with cloud mailboxes filters spam, malware, phishing and spoofing by default. Defender for Office 365 Plan 1 adds impersonation protection, link checking and attachment sandboxing. It’s included in Business Premium and E3, and sold as an add-on for other plans.
Is a secure email gateway the same as a spam filter?
A spam filter is one part of it. A gateway also scans for malware, checks links, watches for spoofing and impersonation, and often controls outgoing mail. Microsoft’s and Google’s built-in protection covers all of that at a basic level.
Do I need a secure email gateway with Google Workspace?
Usually not for a small firm whose mail all lives in Workspace. Gmail’s built-in filtering is on for every account. A separate layer makes sense if a contract or regulator requires a control Google doesn’t provide, or if mail your filter passed has already cost you.
Can a secure email gateway stop business email compromise?
It stops some of it, especially spoofed senders and lookalike domains it has seen before. It struggles with requests from a real account that’s been taken over, because nothing in the message is technically wrong. A call to a number you already have stops what the filter can’t.